Sponsored

Lucid Air over 500 miles on a charge!

ajdelange

Well-Known Member
First Name
A. J.
Joined
Aug 1, 2019
Threads
9
Messages
2,883
Reaction score
2,317
Location
Virginia/Quebec
Vehicles
Tesla XLR+2019, Lexus, Landcruiser, R1T
Occupation
EE Retired
Peter Rawlinson led the chassis and suspension engineering of the Model S. (In fact, Musk is now claiming that's about all he did, despite plenty of evidence to the contrary.) I'm pretty sure he understood the difference between beam and torsional stiffness, and I'm pretty sure he knew the Model S would not have a longitudinal driveshaft. Yet he found the penalty the hatchback opening exacted on frame stiffness to be problematic enough to take steps to avoid it in the Air.
I afraid this is getting too personal but I did not mean to imply that Peter Rawlinson confused torsional and beam stiffness but rather that you do. How would a hatch back interfere with torsional stiffness? It's behind where the torques are applied. It might very well have an effect on beam stiffness though. Now the gull wing doors did present a were a torsional stiffness problem and so the roof in their vicinity had to be beefed up.
Sponsored

 

electruck

Well-Known Member
Joined
Oct 6, 2019
Threads
69
Messages
3,495
Reaction score
6,455
Location
Dallas, TX
Vehicles
2023 Rivian R1S
First off, I don't think you are qualified to lecture engineers on the meaning of engineering terminology.

Second, the Teslas have adaptive suspensions. They sense the road, predict the necessary correction to wheel position and activate sensors to "daintily" raise or lower the individual wheels to smooth the ride and keep the vehicle level.


IMG_1407.jpg


Note the graphical display of the control currents to the actuators trailing each wheel. Don't ask me to explain what the actuators are, what they do or what the adaptive control algorithms behind them are because I don't know but I'd bet more than one beer that fuzzy logic is involved. Note that the 4 wheels are at different heights as is necessary to keep the car level on my uneven gravel driveway. Note also that each wheel is equipped with its own accellerometer.
Nope, sorry. You are definitely wrong about the capabilities here. The Tesla does have air suspension and can adjust the ride height at each corner to level the vehicle and adjust aero effects but this is not at all the same as being able to control the position of the tires relative to the road.

The "adaptive" part of the suspension is the damping. They can vary the amount of damping in both compression and rebound to control the spring force. This is usually done via changes in hydraulic valving. Again, this also does nothing to raise or lower the tires.
 

electruck

Well-Known Member
Joined
Oct 6, 2019
Threads
69
Messages
3,495
Reaction score
6,455
Location
Dallas, TX
Vehicles
2023 Rivian R1S
Lucid is using what it calls "semi-active dampers". They have not explained the term, and I suspect it's some version of an adaptive system.
Here is how Tenneco defines "semi-active dampers": It is a system that continuously adjusts damping levels according to road conditions and vehicle dynamics. This is a pretty common setup on many vehicles today, especially luxury and sport vehicles such as your R8s.
 

Hmp10

Well-Known Member
Joined
Mar 7, 2019
Threads
5
Messages
629
Reaction score
542
Location
Naples, FL
Vehicles
2015 Tesla Model S P90D; 2018 Honda Odyssey
Note the graphical display of the control currents to the actuators trailing each wheel.
That graphical display twice refers to "damping" and nowhere to wheel positioning. I see no mention of actuators. In fact, "compression" and "rebound" are damping terms. How do you know those control currents aren't being used to adjust damping rather than actually to raise or lower the wheels actively?

(Sorry, electruck, I posted this before your post #137 came up. But I left mine up because I'd still like to see ajdelange's answer.)
 

ajdelange

Well-Known Member
First Name
A. J.
Joined
Aug 1, 2019
Threads
9
Messages
2,883
Reaction score
2,317
Location
Virginia/Quebec
Vehicles
Tesla XLR+2019, Lexus, Landcruiser, R1T
Occupation
EE Retired
AJ, you've clearly stepped out of your wheelhouse here. Torsional stiffness is still very much critical to optimal suspension operation. The chassis is essentially a 5th spring in the suspension... and an undamped one at that.
Let me first readily admit that I am an electrical guy very much more than I am a mechanical guy. I recognize that torsional stiffness could be a factor in way a car handles but without doing a finite element analysis of the Tesla frame, which I can't do, I can observe that if 39,000 Nm/° is good enough for a Formula 1 car and that many cars are on the road with 10k or less that 19 - 22k are probably adequate for the S and X. I can also, from basic physics, see that torsional stiffness is more important where the motor mounts try to twist the frame (longitudinal) than where they don't (transverse). I can also see from basic physics that if and adaptive suspension reduces road irregularity induced torques that it is less important. Thus I don't think I've said anything unreasonable but invite further discussion.
 

Sponsored

ajdelange

Well-Known Member
First Name
A. J.
Joined
Aug 1, 2019
Threads
9
Messages
2,883
Reaction score
2,317
Location
Virginia/Quebec
Vehicles
Tesla XLR+2019, Lexus, Landcruiser, R1T
Occupation
EE Retired
Nope, sorry. You are definitely wrong about the capabilities here. The Tesla does have air suspension and can adjust the ride height at each corner to level the vehicle and adjust aero effects but this is not at all the same as being able to control the position of the tires relative to the road.
OK. The ride height changes hundreds of times per second. How do you interpret that?

The "adaptive" part of the suspension is the damping.
Seems to me the ride height part of the system is adapting too.

They can vary the amount of damping in both compression and rebound to control the spring force. This is usually done via changes in hydraulic valving. Again, this also does nothing to raise or lower the tires.
Obviously the damping is being adjusted too. I'm not much concerned with how it is usually done. I am concerned about how it is done in the Tesla.
 

ajdelange

Well-Known Member
First Name
A. J.
Joined
Aug 1, 2019
Threads
9
Messages
2,883
Reaction score
2,317
Location
Virginia/Quebec
Vehicles
Tesla XLR+2019, Lexus, Landcruiser, R1T
Occupation
EE Retired
That graphical display twice refers to "damping" and nowhere to wheel positioning. What does I see no mention of actuators. In fact, "compression" and "rebound" are damping terms. How do you know those control currents aren't being used to adjust damping rather than actually to raise or lower the wheels actively?
One of the big differences between me and you guys is that I am smart enough to know what I don't know and that is a whole lot about how this system actually works. Given the terms I believe the graphs probably are of the damping parameters. But what I see that you didn't is a number that displays ride height and acceleration for each wheel. As these numbers change continuously and rapidly I assume that the ride height is being adjusted dynamically according to driving conditions. I also know full well there is a servo in the system that uses compressed air to raise and lower the car. I'm guessing that it is controlled dynamically but perhaps it isn't. You assure me that it isn't. How do you know that?
 

electruck

Well-Known Member
Joined
Oct 6, 2019
Threads
69
Messages
3,495
Reaction score
6,455
Location
Dallas, TX
Vehicles
2023 Rivian R1S
EDIT: AJ, everything you state about the Tesla in post #142 above is absolutely correct. However...

you seem to be missing a very critical distinction here. Hmp and I aren't saying the Tesla doesn't adjust the ride height (positioning of the body relative to the ground), we're saying it can't lift its tires off the ground. The Tesla absolutely can not lift its tire off the ground to clear an obstacle. Period. End of discussion.
 
Last edited:

ajdelange

Well-Known Member
First Name
A. J.
Joined
Aug 1, 2019
Threads
9
Messages
2,883
Reaction score
2,317
Location
Virginia/Quebec
Vehicles
Tesla XLR+2019, Lexus, Landcruiser, R1T
Occupation
EE Retired
I don't think I'm missing anything. I never meant to say that it lifted the tyres off the ground. Why would it do that? I have to take some credit for causing confusion on your parts by my attempt to be cute with "daintily lifts". What I believe is going on, but again this is only surmise from what I abserve (and one Sandy Munroe video) it that the actuator is a pneumatic cylinder with an outlet valve to ambient and an inlet valve connected to a high pressure reservoir. Let each wheel bear weight equal to
w = m*g equal to 1/4 the weight of the vehicle where m is 1/4 the mass of the car and g the acceleration due to gravity. Model the cylinder as a piston of area A. Let the piston be displaced x. Assume air to be an ideal gas. Then P*V = (F/A)*A*x = n*R*T where P is the pressure in the cylinder equal to the force on the piston divided by its area, V is the volume of the gas contained by the piston, n the number of moles of gas contained, R the gas constant and T the Kelvin temperature.
Then F*x = n*R*T where F nominally equal w, 1/4 the weight of the car. The displacement of the piston, x = n*R*T/w is controllable by increasing or decreasing n i.e. by letting gas out or putting gas into the cylinder. Thus by so doing we can raise or lower the car. Statically or dynamically. Now if we roll over a bump with one wheel the car will be accelerated upward. This will be sensed by the accelerometer on that wheel with an acceleration of a corresponding to a force of a/m. Thus the upward force on the piston becomes (w + a/m) which is (w + a/m)/w larger than before. To cancel the extra upward force all we need do is reduce n to n*w/(w + a/m) by letting gas out. Letting gas out obviously causes x to decrease allowing the wheel to follow the bump (it goes up closer to the car body). The work that would have gone into raising the car by the height of the bump gets dissipated as PV work when the gas is released. If a wheel goes into a rut the process is the same except that gas is admitted to the cylinder so that the piston moves out raising the car relative to the wheel. PV work is done on the car to raise it. The control currents we see in the display are those to the two valves that admit and release air and the inputs (PVs) to the controllers are the accelerometer measurements. The control system tries to keep them 0. What I think you fail to appreciate is the sophistication of the control system algorithms. They can modulate the valves hundreds if not thousands of times per second.

Now if you like you can work through the math above and pretty easily convince yourself that the cylinder behaves like a Hookian spring with constat n*r*T/x0 in which x0 is the nominal piston posiition and the Hookian behaviour relative to that position. And you can probably equally easily show that the rate at which gas is let out of the cylinder corresponds to the lossiness of a dashpot. This might allow you to cast the system in terms of the more classical shock absorber system. But to it we must add the additional degrees of freedom granted by the ability to raise and lower at will and the very rapid adaptation possibilities.

And as it is quite successful we conclude that no torsional forces are applied to the frame so that it does not matter, from this POV at least, how stiff the frame is. Well I suppose it does as the system cannot take out all the challenges of these Quebec unsurfaced roads.
 

thrill

Well-Known Member
First Name
billy
Joined
Jan 14, 2020
Threads
19
Messages
767
Reaction score
1,538
Location
South Carolina
Vehicles
i3s, (r1t)
I don't think I'm missing anything. I never meant to say that it lifted the tyres off the ground. Why would it do that? ...
At 1:40 in this video you'll see an Audi active supension lift the wheel above a pothole. For this to be most effective a rigid frame is needed or the spring effect of the vehicle's mass will cause the lifted wheel to dip sooner than it would otherwise There's other videos on the subject.

 
Last edited:

Sponsored

electruck

Well-Known Member
Joined
Oct 6, 2019
Threads
69
Messages
3,495
Reaction score
6,455
Location
Dallas, TX
Vehicles
2023 Rivian R1S
I never meant to say that it lifted the tyres off the ground.
Never the less, you did.

Well I suppose it does as the system cannot take out all the challenges of these Quebec unsurfaced roads.
So what you are saying here is that while it all sounds good in theory, you may have made a few too many assumptions and overlooked a few minor details that differentiate the observed behavior from that predicted by your simple model. Got it.
 

ajdelange

Well-Known Member
First Name
A. J.
Joined
Aug 1, 2019
Threads
9
Messages
2,883
Reaction score
2,317
Location
Virginia/Quebec
Vehicles
Tesla XLR+2019, Lexus, Landcruiser, R1T
Occupation
EE Retired
This is what I said"... car daintily just picks up the tire so there is no shock..." Where does that say anything about picking it up to the extent that it is off the ground? What I meant was that it picked it up to the point where the shock is not transmitted to the car. That is the point where the force in excess of w is neutralized. That is, as far as I can see, what happens. The accelerometers all read near 0 even on a bumpy road. The distance to ground readings vary. Now if I have learned anything from dealing with the internet I should have learned that one must be very literal. I could have said it better.


I have to make assumptions. All I have are my observations and Sandy Munroe's presentation of the actuator. I don't have any of Tesla's design review materials. What I presented is a model that fits the observations. That model is intended to give you (and me and anyone who looks at what I wrote) some insight. Perhaps you are not familiar with the general concept of mathematical modeling. I haven't observed any differences between the observed behaviour and what the model predicts. Thus it is, at first glance, a reasonable model even though air is not an ideal gas.

If you have an another model, let's hear about it.
 
Last edited:

jjwolf120

Well-Known Member
First Name
John
Joined
Feb 25, 2020
Threads
1
Messages
781
Reaction score
1,121
Location
Arcadia
Vehicles
Rivian R1S
Occupation
TPA
Last edited:

electruck

Well-Known Member
Joined
Oct 6, 2019
Threads
69
Messages
3,495
Reaction score
6,455
Location
Dallas, TX
Vehicles
2023 Rivian R1S
This is what I said"... car daintily just picks up the tire so there is no shock..." Where does that say anything about picking it up to the extent that it is off the ground? What I meant was that it picked it up to the point where the shock is not transmitted to the car.
You still don't get it. I'm not merely questioning how far the suspension "picks up the tire". Since the suspension can not apply any upward force to the wheel, it can not "pick it up" at all. You can increase the spring rate by applying more gas pressure and that will force the wheel down away from the chassis (or the chassis up if the wheel is vertically fixed). You can also reduce the spring rate by reducing the gas pressure but that does not "pick up" the wheel. This simply reduces the force opposing a wheel being lifted by the road (or lowers the chassis if the wheel is vertically fixed).

I haven't observed any differences between the observed behaviour and what the model predicts.
Oh but you have. In one sentence you state that no forces are applied to the frame and then you immediately acknowledge that this is not true based on your observed behavior over the roads of Quebec.

As to your mathematical model, you seem to be assuming infinite reservoirs of high pressure gas capable of instantly transferring whatever mass of gas is needed to achieve ideal suspension behavior. If you give it a little thought, I think you will see how the transfer of gas in and out might start to become a limiting factor in the responsiveness of the air spring (observe letting the air out of a balloon, it takes a finite amount of time). Now I am not stating that your model is way off base or unable to give some insight into what is possible. I am also not stating that this type of suspension can't achieve phenomenal results compared to the classic coil spring and shock suspension. I am merely pointing out that you have not modeled the complete system and should therefore be cautious in drawing conclusions. Where I really feel compelled to call you out though is when you make statements about the wheel being "picked up" by the suspension or that your model predicts that no force should be applied to the frame.

But enough of this, I have better things to do.
 

ajdelange

Well-Known Member
First Name
A. J.
Joined
Aug 1, 2019
Threads
9
Messages
2,883
Reaction score
2,317
Location
Virginia/Quebec
Vehicles
Tesla XLR+2019, Lexus, Landcruiser, R1T
Occupation
EE Retired
As to your mathematical model, you seem to be assuming infinite reservoirs of high pressure gas capable of instantly transferring whatever mass of gas is needed to achieve ideal suspension behavior.
That's a very good observation. The way the thing works is you are saiing along and a wheel hits a bump. This couples energy into the piston. We want that energy absorbed by the gas, not the car. If we keep both valves closed, the piston moves up, the gas compresses, and the temperature goes up. The gas has absorbed energy n*R*(T2 - T1). If the energy input from the ground is more than that the remainder gets coupled into the car and the accelerometer detects this. If it above some threshold then the controller opens the exhaust valve and ∆n moles of gas escape. This disposes of enegy ∆n*R*T (where T is an average temperature between T2 and T1). By controlling the rate at which n is being removed the controller is able to keep the accelerometer reading near 0.

So far you should have been able to follow. From here on I expect you may have some trouble. Assuming the system works the car rides level. Its potential energy stays the same. Using it as a frame of reference what do we see? A fixed body with things attached to it that are going up and down. When a wheel goes up by h the ground did m*g*h work on the system and as the car itself didn't move m is the mass of the wheel and all the work was done on it. Thus the amount of energy which must be disposed of by releasing gas is m*g*h. When we come down off the bump the gas removed must be replaced. Rather than thinking in terms of the number of moles of gas this is think in terms of replacing however much gas is required to restore m*g*h so that the cylinder is in the same state it was before it hit the bump. From this we can calculate that if a wheel weighs 40 kg and a bump is 1 cm high that the compressor will have to restore 40*9.8*.01 = 3.92J worth of air. Supposing the road has one of these every 4 inches and that I am driving 60 mph then the compressor would have to supply of 3*5280*4*40*9.8*.01/3600 = 68.992 W which is of no consequence but what is of consequence is that this would require 68.992 Wh/mi to be added to the other vehicle loads. This, of course, assumes that the entire job is done via the release/replace mechanism. But the message is clear: driving on rough roads with the pneumatic suspension set to highest performance is going to cost you range. No surprise there.

The implied questions about loop bandwidth are good ones too. Tesla's big plus here is that with the incredible onboard computing power they have available controiler loop bandwidths are easily in the high tens of Hz.


I am merely pointing out that you have not modeled the complete system and should therefore be cautious in drawing conclusions.
You are preaching to the choir there. I guess you are not familiar with the concept of modeling. Were you it would not be necessary to point that out.

Where I really feel compelled to call you out though is when you make statements about the wheel being "picked up" by the suspension or that your model predicts that no force should be applied to the frame.
I'm afraid you haven't the perspective to understand the wheel being picked up comment. As to no force being applied to the frame: when the accelerometers all read 0 no torsional force is being applied. The acclerometers all read 0 whenever the system is withing its operating region. If the road gets too rough it cannot control, the accelerometers do read appreciable numbers and there are, depeding on their relative readings, torsional forces.
Sponsored

 
Last edited:
 




Top